Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $X$ be (Edit: a closed convex subset of ) the unit sphere $Y=\{x\in \ell^2: \|x\|=1\}$ in $\ell^2$ with the great circle (geodesic) metric. (Edit: Suppose the diameter of $X$ is less than $\pi/2$.) Is it true that every decreasing sequence of nonempty closed convex sets in $X$ has a nonempty intersection? (A set $S$ is convex in $X$ if for every $x,y\in S$ the geodesic path between $x,y$ is contained in $S$.)

(I edited my original question.)

share|cite|improve this question
Please make your definition of a convex set more clearer. – Hassan Muhammad Oct 31 '11 at 16:59
@Muhammad. If $x\neq -y$, this means $x,y\in S$ implies $\cos(t)x+\sin(t)u \in S$ for $t$ between $0$ and $\cos^{-1}(x,y)$, where $u$ is the unit vector in the direction of $y-(x,y)x$, $(x,y)$ being the inner product of $x,y$. – TCL Oct 31 '11 at 18:00
up vote 3 down vote accepted

No. For example, let $A_n$ be the subset of $X$ consisting of vectors that are zero in the first $n$ co-ordinates.

EDIT: this assumes that when $x$ and $y$ are antipodal, convexity of $S$ containing $x$, $y$ only requires that at least one of the great-circle paths is contained in $S$. If it requires all of them, then the $A_n$ are not convex. t.b. points out in the comments that in this case we can set $A_n$ to consist of all vectors in $X$ that are zero in the first $n$ co-ordinates and non-negative in the remainder.

share|cite|improve this answer
Surely $A_n$ is not convex? For instance, $A_1$ contains $(0,1,0,0,\dots)$ and $(0,-1,0,0,\dots)$ but not $(0,0,0,\dots)$. – Nate Eldredge Oct 31 '11 at 14:12
@Nate: "convex" in this question means as a subset of the unit sphere, with respect to the geodesic metric. It's not yet clear how convexity works with antipodal points though. – Chris Eagle Oct 31 '11 at 14:17
Ah, I see. Thanks. – Nate Eldredge Oct 31 '11 at 14:18
Chris: I think you can fix your example by requiring that all coordinates are non-negative. Then you avoid antipodal points. – t.b. Oct 31 '11 at 14:19
@Eagle. I edited my question. Thanks. – TCL Oct 31 '11 at 14:26

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.